In a previous paper it was shown that persons with a normal heart and patients with cardiac disease behaved differently when subjected to rebreathing experiments.1 When the oxygen in the system reached a concentration ranging between 16 and 13 per cent, the subject began to use less of the gas in the chamber than in the period immediately preceding. Subsequently his use of the gas progressively increased. The patient with cardiac disease, on the other hand, required increasing amounts of gas, varying according to the severity of his disability. The various compensatory mechanisms2 invoked during rebreathing experiments have been considered and evaluated (hyperventilation, increased lung volume, increased number of circulating red cells and increased cardiac output). Of these, cardiac output is considered of paramount importance. This mechanism is known to manifest itself at an oxygen concentration of approximately 14 per cent in normal subjects. This compensation is assumed